If a string rule specifies binary or text file data explicitly, make sure

that the data in the file matches the type before we do any other tests.
This avoids having weak magic types such as msdos, match against text files.

diff --git a/src/ascmagic.c b/src/ascmagic.c
index 67a75dfe283d02d98c6071c4f7cf505433e02a17..f6da80ac3aeed496fc7c8841bb5307ac5f3f3d77 100644 (file)
--- a/src/ascmagic.c
+++ b/src/ascmagic.c
@@ -35,7 +35,7 @@
 #include "file.h"
 
 #ifndef        lint
-FILE_RCSID("@(#)$File: ascmagic.c,v 1.80 2011/03/12 18:13:54 rrt Exp $")
+FILE_RCSID("@(#)$File: ascmagic.c,v 1.81 2011/03/15 22:16:29 christos Exp $")
 #endif /* lint */
 
 #include "magic.h"
@@ -70,7 +70,8 @@ trim_nuls(const unsigned char *buf, size_t nbytes)
 }
 
 protected int
-file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
+file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes,
+       int text)
 {
        unichar *ubuf = NULL;
        size_t ulen;
@@ -93,7 +94,7 @@ file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
        }
 
        rv = file_ascmagic_with_encoding(ms, buf, nbytes, ubuf, ulen, code,
-           type);
+           type, text);
 
  done:
        if (ubuf)
@@ -105,7 +106,7 @@ file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
 protected int
 file_ascmagic_with_encoding(struct magic_set *ms, const unsigned char *buf,
     size_t nbytes, unichar *ubuf, size_t ulen, const char *code,
-    const char *type)
+    const char *type, int text)
 {
        unsigned char *utf8_buf = NULL, *utf8_end;
        size_t mlen, i;
@@ -153,7 +154,7 @@ file_ascmagic_with_encoding(struct magic_set *ms, const unsigned char *buf,
                    == NULL)
                        goto done;
                if ((rv = file_softmagic(ms, utf8_buf,
-                   (size_t)(utf8_end - utf8_buf), TEXTTEST)) != 0)
+                   (size_t)(utf8_end - utf8_buf), TEXTTEST, text)) != 0)
                        goto subtype_identified;
                else
                        rv = -1;

diff --git a/src/file.h b/src/file.h
index 439e291d869ca49b9c90754d91ac02349e5e9c86..610db95b4cfc88d3710ca6ddada15b83356fdb64 100644 (file)
--- a/src/file.h
+++ b/src/file.h
@@ -27,7 +27,7 @@
  */
 /*
  * file.h - definitions for file(1) program
- * @(#)$File: file.h,v 1.133 2011/05/13 22:15:40 christos Exp $
+ * @(#)$File: file.h,v 1.134 2011/09/16 21:23:59 christos Exp $
  */
 
 #ifndef __file_h__
@@ -405,15 +405,16 @@ protected int file_trycdf(struct magic_set *, int, const unsigned char *,
 protected int file_zmagic(struct magic_set *, int, const char *,
     const unsigned char *, size_t);
 #endif
-protected int file_ascmagic(struct magic_set *, const unsigned char *, size_t);
+protected int file_ascmagic(struct magic_set *, const unsigned char *, size_t,
+    int);
 protected int file_ascmagic_with_encoding(struct magic_set *,
     const unsigned char *, size_t, unichar *, size_t, const char *,
-    const char *);
+    const char *, int);
 protected int file_encoding(struct magic_set *, const unsigned char *, size_t,
     unichar **, size_t *, const char **, const char **, const char **);
 protected int file_is_tar(struct magic_set *, const unsigned char *, size_t);
 protected int file_softmagic(struct magic_set *, const unsigned char *, size_t,
-    int);
+    int, int);
 protected struct mlist *file_apprentice(struct magic_set *, const char *, int);
 protected uint64_t file_signextend(struct magic_set *, struct magic *,
     uint64_t);

diff --git a/src/funcs.c b/src/funcs.c
index 71257b05ccba386a2e190a84b1e9fa63f72c229c..026f472eb224cf405c37c539962ffdc581ca3856 100644 (file)
--- a/src/funcs.c
+++ b/src/funcs.c
@@ -27,7 +27,7 @@
 #include "file.h"
 
 #ifndef        lint
-FILE_RCSID("@(#)$File: funcs.c,v 1.56 2011/02/03 01:43:33 christos Exp $")
+FILE_RCSID("@(#)$File: funcs.c,v 1.57 2011/05/11 01:02:41 christos Exp $")
 #endif /* lint */
 
 #include "magic.h"
@@ -228,7 +228,8 @@ file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unu
 
        /* try soft magic tests */
        if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
-               if ((m = file_softmagic(ms, ubuf, nb, BINTEST)) != 0) {
+               if ((m = file_softmagic(ms, ubuf, nb, BINTEST,
+                   looks_text)) != 0) {
                        if ((ms->flags & MAGIC_DEBUG) != 0)
                                (void)fprintf(stderr, "softmagic %d\n", m);
 #ifdef BUILTIN_ELF
@@ -255,7 +256,7 @@ file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unu
        /* try text properties (and possibly text tokens) */
        if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
 
-               if ((m = file_ascmagic(ms, ubuf, nb)) != 0) {
+               if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
                        if ((ms->flags & MAGIC_DEBUG) != 0)
                                (void)fprintf(stderr, "ascmagic %d\n", m);
                        goto done;
@@ -265,7 +266,8 @@ file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unu
                if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
                        if (looks_text == 0)
                                if ((m = file_ascmagic_with_encoding( ms, ubuf,
-                                   nb, u8buf, ulen, code, type)) != 0) {
+                                   nb, u8buf, ulen, code, type, looks_text))
+                                   != 0) {
                                        if ((ms->flags & MAGIC_DEBUG) != 0)
                                                (void)fprintf(stderr,
                                                    "ascmagic/enc %d\n", m);

diff --git a/src/softmagic.c b/src/softmagic.c
index f25f2e5b166496788a3452081523add37071783e..624e57203b21f6f55722313e36d920f2584a95ad 100644 (file)
--- a/src/softmagic.c
+++ b/src/softmagic.c
@@ -32,7 +32,7 @@
 #include "file.h"
 
 #ifndef        lint
-FILE_RCSID("@(#)$File: softmagic.c,v 1.144 2011/01/07 23:22:28 rrt Exp $")
+FILE_RCSID("@(#)$File: softmagic.c,v 1.145 2011/05/13 22:15:40 christos Exp $")
 #endif /* lint */
 
 #include "magic.h"
@@ -43,9 +43,9 @@ FILE_RCSID("@(#)$File: softmagic.c,v 1.144 2011/01/07 23:22:28 rrt Exp $")
 
 
 private int match(struct magic_set *, struct magic *, uint32_t,
-    const unsigned char *, size_t, int);
+    const unsigned char *, size_t, int, int);
 private int mget(struct magic_set *, const unsigned char *,
-    struct magic *, size_t, unsigned int);
+    struct magic *, size_t, unsigned int, int);
 private int magiccheck(struct magic_set *, struct magic *);
 private int32_t mprint(struct magic_set *, struct magic *);
 private int32_t moffset(struct magic_set *, struct magic *);
@@ -66,12 +66,14 @@ private void cvt_64(union VALUETYPE *, const struct magic *);
  */
 /*ARGSUSED1*/          /* nbytes passed for regularity, maybe need later */
 protected int
-file_softmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes, int mode)
+file_softmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes,
+    int mode, int text)
 {
        struct mlist *ml;
        int rv;
        for (ml = ms->mlist->next; ml != ms->mlist; ml = ml->next)
-               if ((rv = match(ms, ml->magic, ml->nmagic, buf, nbytes, mode)) != 0)
+               if ((rv = match(ms, ml->magic, ml->nmagic, buf, nbytes, mode,
+                   text)) != 0)
                        return rv;
 
        return 0;
@@ -106,7 +108,7 @@ file_softmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes, in
  */
 private int
 match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
-    const unsigned char *s, size_t nbytes, int mode)
+    const unsigned char *s, size_t nbytes, int mode, int text)
 {
        uint32_t magindex = 0;
        unsigned int cont_level = 0;
@@ -115,6 +117,7 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
        int firstline = 1; /* a flag to print X\n  X\n- X */
        int printed_something = 0;
        int print = (ms->flags & (MAGIC_MIME|MAGIC_APPLE)) == 0;
+       int wrong_type = 0;
 
        if (file_check_mem(ms, cont_level) == -1)
                return -1;
@@ -123,7 +126,13 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
                int flush = 0;
                struct magic *m = &magic[magindex];
 
-               if ((m->flag & mode) != mode) {
+               if (IS_STRING(m->type)) {
+                       if ((text && (m->str_flags & STRING_BINTEST)) ||
+                           (!text && m->str_flags & STRING_TEXTTEST))
+                               wrong_type = 1;
+               }
+                       
+               if (wrong_type || (m->flag & mode) != mode) {
                        /* Skip sub-tests */
                        while (magic[magindex + 1].cont_level != 0 &&
                               ++magindex < nmagic)
@@ -135,7 +144,7 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
                ms->line = m->lineno;
 
                /* if main entry matches, print it... */
-               switch (mget(ms, s, m, nbytes, cont_level)) {
+               switch (mget(ms, s, m, nbytes, cont_level, text)) {
                case -1:
                        return -1;
                case 0:
@@ -218,7 +227,7 @@ match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
                                        continue;
                        }
 #endif
-                       switch (mget(ms, s, m, nbytes, cont_level)) {
+                       switch (mget(ms, s, m, nbytes, cont_level, text)) {
                        case -1:
                                return -1;
                        case 0:
@@ -1013,7 +1022,7 @@ mcopy(struct magic_set *ms, union VALUETYPE *p, int type, int indir,
 
 private int
 mget(struct magic_set *ms, const unsigned char *s,
-    struct magic *m, size_t nbytes, unsigned int cont_level)
+    struct magic *m, size_t nbytes, unsigned int cont_level, int text)
 {
        uint32_t offset = ms->offset;
        uint32_t count = m->str_range;
@@ -1578,7 +1587,7 @@ mget(struct magic_set *ms, const unsigned char *s,
                if (nbytes < offset)
                        return 0;
                return file_softmagic(ms, s + offset, nbytes - offset,
-                   BINTEST);
+                   BINTEST, text);
 
        case FILE_DEFAULT:      /* nothing to check */
        default: